视觉导航AGV多路径快速检测算法研究

The research of fast multi-path detection algorithm for visual AGV

为了实现视觉导航AGV在分叉路下的自动导引,提出了一种多路径快速检测算法。首先对图像进行灰度化、平滑、边缘检测等操作得到梯度图像,然后结合多路径图像特征,利用行扫描方式提取各路径边缘坐标,最后利用改进的增量式最小二乘法拟合路径中心线。实验表明,该算法运算简单、实时性强,适用于基于嵌入式系统开发的视觉导航AGV。

In order to achieve the automated guide for visual AGV under thefork,a fast multi-path detection algorithm is designed in this paper. First the gradient imageisgained by processes of graying, smoothing, and edge detection. Thenwith multi-path image features, path＇s edge is extracted by line scanning. Finally the center line of path is fitted by improved incremental least square method. The experiment demonstrates that the algorithm with simple operation and hard real-time is suitable for visual AGV based on embedded system.